With great powers comes great responsibility. We have access to data but our ability to evaluate it is lacking. Like a 90 year old using photoshop or a 16 year fresh from the DMV driving a Ferrari. It seems like I run across an article, either main stream media or “trader” alike that attempts to prove a hypothesis using two data points. Time and something else.
Here is what is known, the only thing these types of charts are good for is confirming bias and robo producing bullshit content. Charts are a poor mans infographic. The only thing that matters in data, at least at a useful and decision making level, is context.
If you produce or use these charts take some time and learn about overfitting. Here are some google search results.
Here are the assumptions/overfitting that occurs with these types of charts:
What is most important: The publisher is assuming that what is measured on x or y axis is the most important thing. This goes against about everything we know about human motivation. I do not think we always act in our best interest but we all eventually do what we have to do. This means that we all have a pain point or run out of money.
We can actually just stop right there, that is all that matters. If you need further convincing think about how markets change, why a monkey is a better stock picker than you, and why models fail.
So next time you are going to save the world with your analysis or reveal the next monster trade know that there are at least some people who are going to look at your chart and see it as this:
I love charts, I use charts, I will continue to look at charts. But for the love of all that is holy please stop trying to prove something significant with a signal data point. A chart records it does not explain. You are better than that.
We would really appreciate your feedback, if you like, hate, or think we are full of crap. Please leave a comment, a voice mail (312) 725-9121 , email info @ traderhabits (dot) com or twitter, stocktwits, youtube and facebook. Subscribe to Traderhabits by email or to newsletter.